Cooperative caching method for multi-bitrate video based on edge computing

Abstract

The application provides a kind of collaborative caching method of multi-bit rate video based on edge computing, the application studies the MBS caching placement mechanism of multi-bit rate video, and formulates an integer programming problem.Under the constraint of caching capacity, the problem of how to deploy multiple bit rate videos on MES is proposed, with the goal of maximizing user perceived QoE;The application provides a system model, including a network model, a caching model and a QoE function;For a general QoE function, an efficient caching algorithm based on greedy heuristic is designed, and for a linear QoE formula, it is proved that there is an optimal solution, which converts the problem into a monotone submodule maximization problem.Using the video collaborative caching strategy of the application, better performance can be achieved, and the user perceived QoE is improved.

Description

Technical Field

[0001] This invention relates to a collaborative caching method for multi-bitrate video based on edge computing, belonging to the fields of edge computing and video caching technology. Background Technology

[0002] In traditional mobile streaming systems, each mobile user (MU) downloads video independently of the streaming server via a 4G cellular network. However, delivery systems often fail to meet users' emerging ultra-high-definition demands, leading to video rebuffering and reduced Quality of Experience (QoE). With the rapid growth of mobile video traffic, video buffering and delivery issues are likely to become even more prominent. To address this problem, mobile edge computing has emerged as a promising paradigm, providing significant computing and storage resources to radio access networks to improve video delivery quality.

[0003] To improve user experience, mobile edge caching is a suitable approach. By placing video content closer to the user, video buffering and loading times can be significantly reduced. A common solution is to connect a mobile edge server (MES) to the storage capacity of a small cell base station and then cache the video on the MES. This solution enables high-frequency reuse or high-density spatial reuse of cellular spectrum, providing high-rate, low-latency, and low-packet-loss data transmission for video distribution. Previous articles have formulated the relationship between QoE value and bitrate as a linear or concave function, while we propose a caching mechanism for a strictly increasing QoE function. Furthermore, in practical streaming systems, some popular videos often attract a large number of playback requests from mobile users (MUs). In this case, MES caching can help eliminate redundant transmission of popular videos on the backhaul link. That is, ESS caching can reduce the deployment and maintenance costs of the connection between the cellular base station and the core network, thereby improving the performance and efficiency of the entire system. Based on practical considerations and storage capacity limitations, we explore how to deploy multiple bitrate videos on the mobile edge server to maximize MU-aware QoE.

[0004] In view of this, it is indeed necessary to propose a collaborative caching method for multi-bitrate video based on edge computing to solve the above problems. Summary of the Invention

[0005] The purpose of this invention is to provide a collaborative caching method for multi-bitrate video based on edge computing, thereby achieving user-perceived QoE.

[0006] To achieve the above objectives, this invention provides a collaborative caching method for multi-bitrate video based on edge computing, mainly comprising the following steps:

[0007] Step 1: Construct the network model of the system. Consider a hybrid cellular network consisting of one MBS, multiple SBSs, and multiple MUs. MBS represents a macro base station, SBS represents a small base station, and MU represents a mobile user. An edge cache is deployed on each MBS and each SBS, capable of storing partial files to quickly respond to user requests. N = {1, 2, ..., N} represents different sets of SBSs, deployed in different areas, and their coverage areas do not overlap. Each user connects directly to the MBS, and each user only connects to the network of their assigned SBS. U = {1, 2, ..., U} represents the set of users.

[0008] Step 2: Construct a typical caching model for a multi-bit video system. The set of videos available for user request is defined as V = {1, 2, ..., V}. Each video is encoded into multiple resolutions with different bitrates using a constant bitrate coding scheme. All bitrate levels are represented by L = {1, 2, ..., L}. l This indicates the bitrate corresponding to the video. Higher bitrate levels have higher bitrates, i.e., b1. <b2<…<b L Assume all videos have the same constant bit rate in the encoding scheme and the same video duration, i.e., videos with bit rate level L have a bit rate determined by r. L Representing the same size; dividing long videos into segments of the same size to scale to different video size scenarios; over a period of time, assuming the average user demand for each video v∈V is known, using q v This indicates that systems in popular, practical streaming services can effectively learn the probability of video access, using binary variables. When the value of the binary variable is 1, it indicates that the video v with bit rate l is cached on MES n; each MES has a cache with a storage capacity of M to store multiple video copies, assuming M is r l When a user (MU) requests video via a cellular link, assuming the user downloads video with the lowest bitrate via the cellular link, if the user u∈U requests video v∈V and is not within the coverage area of ​​any MES that has cached any copy of video v, then the user u will directly download the video with the lowest bitrate level from the MBS; otherwise, the user u will download video v with the highest bitrate level available at some MES, using N. u This represents the set of MES that can communicate with MU u;

[0009] Step 3: Give the QoE function. QoE represents the quality of experience perceived by users. Assume that the QoE function is only related to the bitrate. When a higher bitrate is selected, a greater QoE is obtained. Formally, let h: {b1, …, b l} → R represent the QoE function. This function is positive and strictly increasing, that is, for any i < j and i, j ∈ L, it can be obtained that h(b i ) < h(b j ). The QoE function is usually expressed as a linear function or a concave function. Here, the fairness function is defined as

[0010]

[0011] Step 4: Formalize the QoE maximization problem for multi-bitrate video and MES cache assistance. In the multi-bitrate video collaborative caching in mobile edge computing, determine the optimization problem to maximize the total QoE perceived by users; Let the set The cache decision set C is a subset of the set P. Let L u,v (C) represent the bitrate selected for the MU u requesting the video v under the cache decision C, where, when x ≥ 1, [x] 1+ = x, otherwise [x] 1+ = 1; The QoE perceived by the MU u requesting the video v is expressed as h(L u,v (C)). According to the total probability formula, the QoE perceived by the MU u is where That is, the total user-perceived QoE of all users under the decision C is represented by That is Finally, formalize the QoE maximization problem for supporting multi-bitrate video and MES cache assistance;

[0012] The objective function is

[0013]

[0014]

[0015] Step 5: Analyze the multi-bitrate video caching problem in edge computing, and design a low-complexity and easy-to-implement algorithm to obtain a sub-optimal solution;

[0016] Step 6: Determine the collaborative placement scheme of multi-bitrate videos on the edge server, and obtain the corresponding strategy with the goal of maximizing the QoE perceived by users;

[0017] Step 7: The server caches the corresponding videos according to the obtained cache placement set and achieves the optimization goal.

[0018] As a further improvement of the present invention, in step 1, a caching scheme is executed on the communication graph formed during the busy period of the streaming system, and the caching solution is...

[0019] As a further improvement of the present invention, in step 5, the proposed user-perceived quality of experience (QoE) is understood as the sum of the utility values ​​of all SBSs in the caching system, which can transform the optimization objective into maximizing the utility of all SBSs; the analysis problem is a knapsack problem, which can be regarded as an example of the well-known multiple-choice knapsack problem, and there exists a set of v represented by {I1, I2, ..., I...} v Each set has L items, that is, the l-th item is in set I. v The corresponding value is a bit rate level b. l A copy of video v is generated, with a revenue of p and a weight of video size r. l Furthermore, corresponding to the cache of the MES is the backpack capacity of M, so the problem is transformed into maximizing the total profit of the items in the backpack under the capacity constraint.

[0020] As a further improvement of the present invention, the problem to be solved by the collaborative placement scheme in step 6 is essentially the classic problem of maximizing the monotonic submodulus function under the constraint of matroid. The files are repeatedly added to the content caching decision set, and each selected file brings the maximum marginal utility value to the set. Finally, when all the cache space is filled with files, the content caching decision set will become the best caching strategy.

[0021] The specific steps of the collaborative cache placement algorithm are as follows:

[0022] Step 6.1: Initialize the placement set and cache placement collection

[0023] Step 6.2: Determine if the placement set P is empty. If it is not empty, calculate the value from P. n The marginal utility of adding each file;

[0024] Step 6.3: Select the cache decision that brings the greatest marginal utility

[0025] Step 6.4: Set P n Decisions that meet the criteria are placed into the corresponding cache placement set C;

[0026] Step 6.5: If Will Remove from C and remove from P n Remove from set P;

[0027] Step 6.6: Repeat steps 6.2 to 6.5 until set P is empty;

[0028] Step 6.7: The final output cache is placed in set C.

[0029] The beneficial effects of this invention are as follows: Compared with the prior art, this invention formalizes the collaborative caching problem of multi-bitrate video based on mobile edge computing, maximizing the QoE perceived by mobile users. First, it proposes a network model in MEC networks, a typical caching model and QoE function in multi-bitrate video. Second, it formalizes the QoE maximization problem of multi-bitrate video and MES caching assistance. To address the challenges posed by the original problem, the proposed user-perceived quality of experience (QoE) is understood as the sum of the utility values ​​of all SBSs in the caching system. Therefore, the optimization objective is transformed into maximizing the utility of all SBSs, leading to the conclusion that the proposed problem is essentially a knapsack problem. The problem is viewed as an example of the well-known multi-choice knapsack problem. A method for video file content caching decision set based on the ACS algorithm is introduced. The core idea is to put the files into an empty set and gradually make the set reach the maximum utility value under the constraint of cache capacity. Attached Figure Description

[0030] Figure 1 This is a schematic diagram of the system model in the mobile edge computing of the present invention.

[0031] Figure 2 This is a flowchart illustrating the collaborative caching method for multi-bitrate video based on edge computing according to the present invention. Detailed Implementation

[0032] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be described in detail below with reference to the accompanying drawings and specific embodiments.

[0033] It should be noted that, in order to avoid obscuring the present invention with unnecessary details, only the structures and / or processing steps closely related to the present invention are shown in the accompanying drawings, while other details that are not closely related to the present invention are omitted.

[0034] Additionally, it should be noted that the terms “comprising,” “including,” or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0035] This embodiment proposes a system model for multi-bitrate video co-caching based on mobile edge computing, as follows: Figure 1As shown; considering the caching methods for multi-bitrate videos, and aiming to maximize the user-perceived QoE, the video caching decision is solved. The specific process of this multi-bitrate video collaborative caching problem in edge computing is as follows. Figure 2 As shown, it includes the following steps:

[0036] Step 1: Describe the system's network model: Consider a hybrid cellular network consisting of a macro base station (MBS), multiple small base stations (SBS), and multiple mobile users (MU). An edge cache is deployed on each MBS and SBS, capable of storing files to quickly respond to user requests. N = {1, 2, ..., N} represents the different sets of SBS, deployed in their respective areas, with no overlap in their coverage areas. U = {1, 2, ..., U} represents the set of users.

[0037] Step 2: Provide a typical caching model for a multi-bit video system: The set of videos available for user request is defined as V = {1, 2, ..., V}. Each video is encoded into multiple resolutions with different bitrates using a constant bitrate (CBR) coding scheme. Let L = {1, 2, ..., L} represent all bitrate levels, b l This indicates the bitrate corresponding to the video. Higher bitrate levels have higher bitrates, i.e., b1. <b2<…<b L Assuming all videos have the same constant bit rate in the encoding scheme and the same video duration, videos with bit rate level L have a bit rate determined by r. L The same size is represented. Over a period of time, assuming the average user demand for each video v∈V is known, q is used. v This indicates that systems in popular, practical streaming services can effectively learn the probability of video access. (Binary variable) When its value is 1, it indicates that video v with bitrate l is cached on MES n. Each MES has a cache with a storage capacity of M to store multiple video copies, assuming M is r l The multiples of. When a MU requests video via a cellular link, we assume the user downloads the video with the lowest bitrate via the cellular link. If user u∈U requests video v∈V and the user is not within the coverage area of ​​any MES that has cached any copy of video v, then user u will download the video with the lowest bitrate level directly from the MBS; otherwise, user u will download the video v with the highest bitrate level available at some MES. N u This represents the set of MES that can communicate with MU u.

[0038] Step 3, Give the QoE function: We assume that QoE is only related to the bitrate. When the selected bitrate is higher, the obtained QoE is also larger. Formally, let h: {b1,…,b l} → R represent the QoE function, which is positive and strictly increasing, that is, for any i < j and i, j ∈ L, h(b i ) < h(b j ). The QoE function is usually expressed as a linear function or a concave function. Here, the fairness function is defined as

[0039]

[0040] Step 4, Formalize the QoE maximization problem for multi-bitrate video and MES caching assistance: In the multi-bitrate video collaborative caching in mobile edge computing, we determine the optimization problem aiming to maximize the total QoE perceived by users. Let the set The caching decision set C is a subset of the set P. Let L u,v (C) denote the bitrate selected for the MU u requesting the video v under the caching decision C, where when x ≥ 1, [x] 1+ = x, otherwise [x] 1+ = 1. Therefore, the QoE perceived by the MU u requesting the video v is expressed as h(L u,v (C)). According to the total probability formula, the QoE perceived by the MU u is where So the total user-perceived QoE of all users under the decision C is represented by That is, Finally, we formalize the QoE maximization problem supporting multi-bitrate video and MES caching assistance.

[0041] The objective function is

[0042]

[0043]

[0044] Step 5, Analyze the multi-bitrate video caching problem in edge computing and design a low-complexity and easy-to-implement algorithm to obtain a sub-optimal solution. The user-perceived quality of experience (QoE) we proposed can be understood as the sum of the utility values of all SBSs in the caching system. Therefore, the optimization goal can be transformed into maximizing the utility of all SBSs. Analyzing the problem is essentially a knapsack problem. The problem can be regarded as an example of the well-known multiple-choice knapsack problem.

[0045] Step 6: Determine the collaborative caching placement scheme for multi-bitrate video on the edge server: A corresponding strategy is derived with the goal of maximizing user-perceived QoE. The specific steps of the collaborative caching placement algorithm are as follows:

[0046] Step 6.1: Initialize the placement set and cache placement collection

[0047] Step 6.2: Determine if the placement set P is empty. If it is not empty, calculate the value from P. n The marginal utility of adding each file.

[0048] Step 6.3: Select the cache decision that brings the greatest marginal utility Step 6.4: Set P n Decisions that meet the criteria are placed into the corresponding cache placement set C. Step 6.5: If... Will Remove from C and remove from P n Remove from set P.

[0049] Step 6.6: Repeat steps 6.2 to 6.5 until set P is empty.

[0050] Step 6.7: The final output cache is placed in set C.

[0051] Step 7: The server places the corresponding videos into a cache set based on the obtained cache and achieves the optimization goal.

[0052] In summary, this invention formalizes the collaborative caching problem for multi-bitrate video based on edge computing, improving the user-perceived quality of experience. First, it establishes the system's network model, caching model, and QoE function. Second, it formalizes the QoE maximization problem for multi-bitrate video and MES-assisted caching. To address the challenges posed by the original problem, the proposed user-perceived quality of experience (QoE) is transformed into maximizing the sum of the utility values ​​of all SBSs in the caching system. An algorithm based on the ACS algorithm for video file content caching decision sets is adopted. Finally, it achieves edge computing-based caching of multi-bitrate video, improving the user-perceived quality of experience.

[0053] The above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A collaborative caching method for multi-bitrate video based on edge computing, characterized in that, The specific steps include the following: Step 1: Construct the network model of the system, considering a hybrid cellular network consisting of one and multiple and multiple Composition, in which, Indicates macro base station, Indicates a small base station. Indicates the user, the and each of the above An edge cache is deployed on each platform, which can store a portion of the files to quickly respond to user requests. Indicates different Sets, different descriptions The collections are deployed in their respective regions, and the different... The coverage areas of the collections do not overlap; each user is directly connected to the collection. Each user only connects to the address they are registered with. The network, Represents a set of users; Step 2: Construct a typical caching model for a multi-bit video system, where the set of videos available for user requests is defined as... Each video is encoded into multiple resolutions with different bit rates using a constant bit rate coding scheme. Represents all bit rate levels. This indicates the bitrate corresponding to the video; higher bitrate levels have higher bitrates, i.e. Assume all videos have the same constant bit rate in the encoding scheme, and the video duration is the same, i.e., the bit rate level is constant. The video has the characteristics of Representing the same size; dividing long videos into segments of the same size to scale to different video size scenarios; assuming each video is known over a period of time. Average user demand, using This indicates that systems in popular, practical streaming services can effectively learn the probability of video access, using binary variables. , When the value of the binary variable is 1, it indicates that it has a bit rate. Video Cached to Above; each Each has a storage capacity of The high-speed cache is used to store multiple video copies, assuming yes Multiples of, when When requesting video via a cellular link, assuming the user is downloading a video with the lowest bitrate via the cellular link, if the user Request video At that time, the user is no longer using the cached video. Any copy of any Within the coverage area, then the user will directly from Download videos with the lowest possible bitrate; otherwise, the user... Download has certain Video at the maximum available bitrate level ,use Indicates the ability to communicative gather; Step 3, give function, This represents the user's perceived quality of experience, assuming... It depends only on the bit rate; when a higher bit rate is selected, the obtained... It is also larger, in form, setting express A function that is positive and strictly increasing, meaning that for any , . and All can be obtained , Functions are typically represented as linear or concave functions; here, the fairness function is defined as follows: ； Step 4: Formalize multi-bit video and Cache assistance Maximizing the user experience in multi-bit video collaborative caching within mobile edge computing is an optimization problem. Total amount; let set Cache decision set It is a set A subset, let Indicating in caching decisions The following is the requested video. of Selected bit rate , among which, when ,otherwise Request video of Perceived Represented as According to the total probability formula, Perceived for ,in That is, all users in the decision-making process Total user perception Depend on It means, that is Finally, formalization supports multi-bitrate video and Cache assistance Maximization problem; The objective function is , ， ； Step 5: Analyze the multi-bitrate video caching problem in edge computing, and design a low-complexity, easy-to-implement algorithm to obtain a suboptimal solution; in Step 5, the proposed user-perceived quality of experience... Understand as all of the caching system The sum of utility values ​​can transform the optimization objective into making all To maximize utility; the analysis problem is a knapsack problem, viewed as an example of the well-known multiple-choice knapsack problem, where there exists a... The set is represented as Each set has Item, i.e., the first Item in set The corresponding value is a bit rate level. Video The revenue from the reproduction is The weight is the size of the video. Furthermore, with the aforementioned The cache corresponds to the backpack capacity. The problem then becomes maximizing the total profit of the items in the knapsack under a capacity constraint. Step 6: Determine a collaborative placement scheme for multi-bitrate video on the edge server to maximize user experience. To obtain the corresponding strategy for the objective; Step 7: The server places the corresponding videos into a cache set based on the obtained cache and achieves the optimization goal.

2. The collaborative caching method for multi-bitrate video based on edge computing according to claim 1, characterized in that, In step 1, a caching scheme is executed on the communication graph formed during the busy period of the streaming system, and the caching solution is implemented.

3. The collaborative caching method for multi-bitrate video based on edge computing according to claim 1, characterized in that, The problem that the collaborative placement scheme in step 6 aims to solve is essentially the classic problem of maximizing the monotonic submodulus function under the constraint of matroid. It involves repeatedly adding files to the content caching decision set, with each selected file bringing the maximum marginal utility value to the set. Finally, when all cache space is filled with files, the content caching decision set becomes the optimal caching strategy. The specific steps of the collaborative cache placement algorithm are as follows: Step 6.1: Initialize the placement set , and cache placement collection ; Step 6.2: Determine the placement set If it is not empty, calculate the value from... The marginal utility of adding each file; Step 6.3: Select the cache decision that brings the greatest marginal utility ; Step 6.4: Set Decisions that meet the criteria are placed into the corresponding cache placement set. middle; Step 6.5: If ,Will from Delete it, and From the set Delete; Step 6.6: Repeat steps 6.2 to 6.5 until set P is empty; Step 6.7: Final output cache placement set .

Images